alextransvideo/AlexVoice-GPU-Fast

🦜 VieNeu-TTS-0.3B

Overview

VieNeu-TTS-0.3B is an ultra-fast, on-device Vietnamese Text-to-Speech (TTS) model with instant voice cloning.

Unlike the original 0.5B version, this 0.3B model is trained from scratch on the VieNeu-TTS-1000h dataset. It is optimized for extreme efficiency, delivering 2x faster inference while maintaining high speech quality.

Voice Cloning: All model variants (including GGUF) support instant voice cloning with just 3-5 seconds of reference audio.

Tác giả: Phạm Nguyễn Ngọc Bảo

☕ Support This Project

Training high-quality TTS models requires significant GPU resources. If you find this model useful, please consider supporting the development:

---

🦜 Voice Cloning Inference

Reference Voice (Speaker Example):

Input Text:

Trên bầu trời xanh thẳm, những đám mây trắng lửng lờ trôi như những chiếc thuyền nhỏ đang lướt nhẹ theo dòng gió. Dưới mặt đất, cánh đồng lúa vàng rực trải dài tới tận chân trời, những bông lúa nghiêng mình theo từng làn gió.

Generated Output (Cloned Voice):

---

🚀 Quick Start (Web UI)

1. Requirements (eSpeak NG)

eSpeak NG is mandatory for phonemization.

• Windows: Download .msi from eSpeak NG Releases.
• macOS: brew install espeak
• Linux: sudo apt install espeak-ng

2. Installation & Run

git clone https://github.com/pnnbao97/VieNeu-TTS.git
cd VieNeu-TTS

# 1. Install uv (if you haven't)
# Windows: powershell -c "irm https://astral.sh/uv/install.ps1 | iex"
# Linux/macOS: curl -LsSf https://astral.sh/uv/install.sh | sh

# 2. Sync environment 
uv sync

# 3. Launch Web UI
uv run gradio_app.py

3. Demo Video

---

📦 Using Python SDK (vieneu)

Install the SDK to integrate VieNeu-TTS-0.3B into your research or applications:

# Windows (Avoid llama-cpp build errors)
pip install vieneu --extra-index-url https://pnnbao97.github.io/llama-cpp-python-v0.3.16/cpu/

# Linux / MacOS
pip install vieneu

Full Features Guide

from vieneu import Vieneu
import os

# Initialization
tts = Vieneu()  # Default: 0.3B-Q4 GGUF for CPU
os.makedirs("outputs", exist_ok=True)

# 1. List preset voices
available_voices = tts.list_preset_voices()
for desc, name in available_voices:
    print(f"   - {desc} (ID: {name})")

# 2. Use specific voice (dynamically select second voice)
if available_voices:
    _, my_voice_id = available_voices[1] if len(available_voices) > 1 else available_voices[0]
    voice_data = tts.get_preset_voice(my_voice_id)
    audio_spec = tts.infer(text="Chào bạn, tôi đang nói bằng giọng của bác sĩ Tuyên.", voice=voice_data)
    tts.save(audio_spec, f"outputs/standard_{my_voice_id}.wav")
    print(f"💾 Saved synthesis to: outputs/standard_{my_voice_id}.wav")

# 3. Standard synthesis (uses default voice)
text = "Xin chào, tôi là VieNeu. Tôi có thể giúp bạn đọc sách, làm chatbot thời gian thực, hoặc thậm chí clone giọng nói của bạn."
audio = tts.infer(text=text)
tts.save(audio, "outputs/standard_output.wav")
print("💾 Saved synthesis to: outputs/standard_output.wav")

# 4. Zero-shot voice cloning
if os.path.exists("examples/audio_ref/example_ngoc_huyen.wav"):
    cloned_audio = tts.infer(
        text="Đây là giọng nói đã được clone thành công từ file mẫu.",
        ref_audio="examples/audio_ref/example_ngoc_huyen.wav",
        ref_text="Tác phẩm dự thi bảo đảm tính khoa học, tính đảng, tính chiến đấu, tính định hướng."
    )
    tts.save(cloned_audio, "outputs/standard_cloned_output.wav")
    print("💾 Saved cloned voice to: outputs/standard_cloned_output.wav")

# 5. Cleanup
tts.close()

Remote Mode (Ultra-Fast with LMDeploy Server)

For maximum speed, deploy a Docker server first, then connect remotely:

Step 1: Deploy Docker Server

docker run --gpus all -p 23333:23333 pnnbao/vieneu-tts:serve --model pnnbao-ump/VieNeu-TTS-0.3B --tunnel

Step 2: Connect from Client

from vieneu import Vieneu
import os

# Configuration
REMOTE_API_BASE = 'http://your-server-ip:23333/v1'  # Or bore.pub:XXXX
REMOTE_MODEL_ID = "pnnbao-ump/VieNeu-TTS-0.3B"

# Initialization (LIGHTWEIGHT - only loads small codec locally)
tts = Vieneu(mode='remote', api_base=REMOTE_API_BASE, model_name=REMOTE_MODEL_ID)
os.makedirs("outputs", exist_ok=True)

# List remote voices
available_voices = tts.list_preset_voices()
for desc, name in available_voices:
    print(f"   - {desc} (ID: {name})")

# Use specific voice
if available_voices:
    _, my_voice_id = available_voices[1]
    voice_data = tts.get_preset_voice(my_voice_id)
    audio_spec = tts.infer(text="Chào bạn, tôi đang nói bằng giọng của bác sĩ Tuyên.", voice=voice_data)
    tts.save(audio_spec, f"outputs/remote_{my_voice_id}.wav")
    print(f"💾 Saved synthesis to: outputs/remote_{my_voice_id}.wav")

# Standard synthesis
text_input = "Chế độ remote giúp tích hợp VieNeu vào ứng dụng Web hoặc App cực nhanh mà không cần GPU tại máy khách."
audio = tts.infer(text=text_input)
tts.save(audio, "outputs/remote_output.wav")
print("💾 Saved remote synthesis to: outputs/remote_output.wav")

# Zero-shot voice cloning (encodes audio locally, sends codes to server)
if os.path.exists("examples/audio_ref/example_ngoc_huyen.wav"):
    cloned_audio = tts.infer(
        text="Đây là giọng nói được clone và xử lý thông qua VieNeu Server.",
        ref_audio="examples/audio_ref/example_ngoc_huyen.wav",
        ref_text="Tác phẩm dự thi bảo đảm tính khoa học, tính đảng, tính chiến đấu, tính định hướng."
    )
    tts.save(cloned_audio, "outputs/remote_cloned_output.wav")
    print("💾 Saved remote cloned voice to: outputs/remote_cloned_output.wav")

---

📋 Reference Voices

File	Gender	Accent	Description
Bình	Male	North	Male voice, North accent
Tuyên	Male	North	Male voice, North accent
Nguyên	Male	South	Male voice, South accent
Hương	Female	North	Female voice, North accent
Ngọc	Female	North	Female voice, North accent
Đoan	Female	South	Female voice, South accent

---

🔬 Model Variants

Model Variant	Format	Optimization	Quality	Speed
VieNeu-TTS-0.3B	PyTorch	GPU (LMDeploy)	⭐⭐⭐⭐⭐	Ultra Fast
VieNeu-TTS-0.3B-q8-gguf	GGUF Q8	CPU	⭐⭐⭐⭐	Fast
VieNeu-TTS-0.3B-q4-gguf	GGUF Q4	CPU / Mobile	⭐⭐⭐	Extreme Speed

---

📑 License

This model is released under the CC BY-NC 4.0 (Creative Commons Attribution-NonCommercial 4.0 International) license.

• Experimental Version: VieNeu-TTS-0.3B is currently in the experimental stage.
• ✅ Free: For students, researchers, and non-profit purposes.
• ⚠️ Commercial/Enterprise: Use for businesses or commercial products is strictly prohibited without prior authorization. Please contact the author (Phạm Nguyễn Ngọc Bảo) for licensing terms (Estimated: 5,000 USD/year - negotiable).

📑 Citation

@misc{vieneutts03b2026,
  title        = {VieNeu-TTS-0.3B: Ultra-Fast Vietnamese Text-to-Speech trained from scratch},
  author       = {Pham Nguyen Ngoc Bao},
  year         = {2026},
  publisher    = {Hugging Face},
  howpublished = {\url{https://huggingface.co/pnnbao-ump/VieNeu-TTS-0.3B}}
}

---

Made with ❤️ for the Vietnamese TTS community